Hi everyone. I recently built a custom belt driven router table using Openbuilds hardware! Started doing a few pen plots to try and dial it in. I’ve got a weird issue on my x axis (which is very similar to a Lead 1515, but belt driven) that I get scale issues. It’s easiest to see on lettering, and circles. The thing is, it repeats the same problems identically between jobs. It does it the same amount every time. Any ideas where I should be looking? Belts are tight, and I feel if I was missing steps I would see some weirdness in the closing out of the “H”. (The stray line on the “E” was caused by me bumping the pen.) Thanks in advance, Craig
'same amount everytime' is probably backlash, check that the grub screw in the pulley is firmly on the flat of the motor shaft. did you pull the belts tight with a spring balance? The only way to know they are tight enough and all the same is with a balance. Pull to 12kg of tension.
Thanks David! So I bought a tension scale and tensioned the x axis to 12kg. Also double checked the grub screws and we are snuggled up. Wheels are tensioned and also looking good. I’m still having what looks like to be backlash issues, with it improving somewhat. Here’s the test: Here’s a pic of the setup: Any ideas where to look next? Or am I better seeking out a lead screw setup on this big rig? Thanks, Craig
Is this with a pen? If so, Is there a chance there is some flexing of the pen, or bending of the tip?
Possibly? I had to adjusted to just dragging, but there could be some play. I’m going to put the router back on and make a cut to see how it goes. I’ll post the results. Thanks
I did a quick cut: so it looks like my problem is still there, along with some ovaled circles and over cut “L”s. So the “H” turned out something like this: I was thinking of enlarging it by a factor of 4. If it was backlash, the difference between the measurements should stay the same. Whereas if it is something weird, it would also scale up by 4. Any other thoughts??
Did the bigger cut (3x) Differences in width on the “H” stayed the same at around .30mm (⬅️ then ➡️) and .41mm (➡️ then ⬅️). So I’m assuming it’s a constant backlash issue. Could it be the fact that I’m belt driving the top gantry instead of the lower gantry? Thanks for letting be bounce ideas off you guys.
oh yes, driving the top is a long way from the force on the bit (allowing more flex), and that X carriage is awfully narrow compared to the height.
Well, I’ll drop the drive down and see what she does! I’m surprised with only a pen as a load that I could get that much flex, but I guess it doesn’t take much to produce .3 mm. Thanks again for the advice!
So I moved the drive gantry to the lower x axis, double belted it, and the backlash is still there! I’m not sure what do to from here. The gantry has a ton of plates, braces and bolts, I’m doubting I’m racking, especially under just the load of dragging a pen. If anything the backlash got worse when I moved it down. Could a bad stepper cause some sort of consistent misstep like this??
I decided to do some long jogs back and forth to see if any issues popped up. I put a clamp on the x axis marking the left side of the x gantry. I then jogged over 1000mm, and then back. It parked itself dead on. I’m at a loss on where to go from here. When I was building this machine, I was more worried about my custom y axis setup and decided to go with a canned solution for the x axis. The funny thing is that the y axis is the one that is preforming perfectly!
jogging with jog keys or Gcode instructions? for this kind of test you must use Gcode instructions like G0 X1000 G0 X0 This ensures that it is travelling at maximum speed. Jogs may be restricted in some way. But, I see in your photo that the pen appears to be held on with rubber bands. There may be some opportunity there for 'backlash' (-:
Hi David! The pen is held with a couple of zip ties, though they kinda do look like rubber bands! The dewalt has a nice little groove machined in that holds a pen pretty dang solidly. I did use the jog function, it makes sense to run the same motion via g code. I’ll do that and report back! Thanks
So running the movements via gcode commands produced a perfect park. Here’s the two shots to compare: If anyone has any ideas where to look I would be most grateful. I just can’t see racking even presenting itself with a pen that is often barely even leaving a mark. -do steppers ever present issues like this? -could another axis cause what seems to be backlash only in the x axis? -do I just go to a lead screw setup and hope that it is belt based??
So an update: I decided to redo the belting on my x axis. I’m using a 15mm wide HTD belt, double belted. I bonded the lower belt directly to the c beam, and printed out some idler pulleys to replace two of the v wheels. Just have the one x axis set up currently to test it out, pretty happy with it!
Thank you! We’ll see how the printed parts hold up in real life use. I might have to them turned from aluminum, if needed.
So I’ve also staring working on converting my y axis’ over to a belt servo style drive. This is the underside of the rail, with another 5 wheels on top. The v slot still has to be filled and the bottom belt bonded, but it seems like we are in business! Excited to see what kind of performance I can get out of this thing! !
